Serial measurements of serum and urine markers of bone metabolism and of forearm bone density (BMD) by dual photon absorptiometry were performed in 22 patients undergoing renal transplantation in 1986. Patients were randomised to immunosuppression with (1) cyclosporin alone (CsA group, n = 10), (2) cyclosporin for 3 months followed by azathioprine-prednisone (CsA/AzP group, n = 3) or (3) long-term azathioprine-prednisone (LT AzP group, n = 9). As no reduction in bone mineral density (BMD) was noted in the first 6 months, groups 2 and 3 were considered together (AzP group, n = 12). Mean +/- SEM BMD fell by 19 +/- 2% at 36 months (n = 19, p < 0.01), with similar reductions seen in the CsA and AzP groups. At 60 months, BMD of the AzP group was 25 +/- 3% below baseline (p < 0.01), while the CsA group were only 5 +/- 4% below baseline (p = NS vs baseline, p < 0.05 vs AzP group). The degree of reduction in BMD over 5 years correlated with total glucocorticoid dose (r = 0.63, p < 0.05), but not with biochemical markers of bone turnover. Serum alkaline phosphatase fell post-transplant in patients treated with AzP, but not in the CsA group. These results demonstrate significant loss of forearm bone mineral with long-term follow-up after renal transplantation, but suggest that patients treated with cyclosporin monotherapy may be at lower risk of this complication.